JChem server stops working

Dear Chemaxon support

A user of our software has reported that the JChem service is stopping unexpectedly.  The only way they find this out is when our end user software reports an error with the JChem connection.

They are using JChem 5.2.6 on a UNIX server.

Do you know why this could be happening?  Also, can you please tell me:

1) Is there a way to check if the service is still running without stopping and restarting the service and is there a way to "keep alive" the service once it is running.  Currently this is just reported by the end user application and so it would be good if this could be highlighted and maybe an alert sent to the users.

2) Are there any additional levels of logging that can be used to identify the problem?  At present no relevant information appears to be written to the logs in the cartridge/logs directory

3) Are there any additional logs stored by the program that we could send you to help us to understand why this occurs

 

We look forward to hearing from you.

Many thanks

Karen

Dear Karen,

An abrupt termination of the server process is most probably due to a JVM crash.

1. You can use the ping-jcserver.sh script in the jchem/cartridge directory to check the availability of the JCC server. Usage instructions are in the script itself in comment.

2. In the case of an abrupt termination of the server, having more detailed logging is not likely to be useful. I'd still appreciate if you could post (or send me by mail) what you have in the jchem/cartridge/logs directory.

3. If my assumption about the problem being the result of a JVM crash is correct, there will be a hs_err...log file (such as hs_err_pid8581.log) in the working directory of the JCC server (which is jchem/cartridge by default). Please, send it to me if you find such a file.

Please, also let me know the Java version and the OS vendor and version.

Regarding a real "keep alive" option (meaning that the server is automatically restarted after a crash), it is not available yet, but is feasible in principle. It requires some problems to be solved such as references to the old server instance becoming "stale" in existing database sessions. (The exact error messages the users in this case have probably received.)

Peter

Hi Peter

Thank you for the additional information and please accept my apologies for the delay in our response.

We've heard back from the user and they have not been able to reproduce the crash so they don't want to pursue this investigation.

Thank you for your assistance in this matter, and if this happens again we'll ask for the log files so that we can provide these with the request for support.

Many thanks for your help.

Karen

Hi Peter

Just to let you know that the user has reported this issue again so I would like to re-open this support request.

I have asked for the information you have requested and I will contact you once this is obtained.  Do you have an email address that I may send this information to?

The user is also asking for a way to proactively monitor the system so that they can find out when this is failing before it is reported by the end user.  We are using JChem 5.2.0.6, and at present this is not something we can alter.

Many thanks

Karen

Hi Karen,

My e-mail address is: pkovacs at chemaxon dot com .

Peter

Please, in addition to what I requested above, have the user send the files found in jchem/cartridge/logs directory as well and have them check the system logs (/var/log/messages on RedHat-based systems) for suspcious entries such as the OOM Killer terminating a process due to insufficient memory.

Thanks

Peter

Thanks Peter

I've asked for the additional files and I'll send them along as soon as I have them.

With Best Wishes

Karen

For the record:

We've recieved JChem Server logs, in which we've found nothing useful. The user reviewed the system logs and found nothing unusual. Java version is 1.6.0_23.

We'ved asked for the stderr/stdout output of the JChem Server if available.

P.